OpenAI's Codex can now watch you work once and repeat the task forever

· Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Software Development & Engineering · Depth: Novice, quick

Summary

OpenAI released the new Record & Replay feature for its Codex app on macOS on June 20, 2026. This functionality allows users to demonstrate a workflow, such as uploading a YouTube video with specific metadata, a thumbnail, and subtitles, just once. Codex then records this process, converting it into a reusable "skill" that can autonomously repeat the task. While the Record & Replay feature is not yet available in the EU, UK, or Switzerland, it requires "Computer Use" to be enabled, which has been available in the EU since June 16. Version 26.616 of the Codex app also introduces bulk actions for Automations history and the ability to hand off threads between local and remote hosts for task continuity. Codex, designed as an AI agent for coding and white-collar automation, is free to download but necessitates a paid ChatGPT account for full utility.

Key takeaway

For Automation Engineers seeking to streamline repetitive digital workflows, OpenAI's Codex Record & Replay feature offers a direct path to automation. You can now demonstrate a complex task once, like publishing a YouTube video or formatting pull requests, and Codex will learn and repeat it autonomously. This significantly reduces the effort traditionally required for scripting or configuring automation rules, allowing you to quickly deploy AI agents for consistent execution of your preferred processes.

Key insights

OpenAI's Codex enables AI agents to learn and automate complex workflows through single-demonstration recording.

Principles

Method

Users perform a task while Codex observes, records the steps and preferences, and then converts this into an executable skill for future automation.

In practice

Topics

Best for: Executive, AI Product Manager, Product Manager, Automation Engineer, AI Engineer, Software Engineer

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.